Agent skill
plan-brainstorm
Explore creative approaches and ideas with external source validation.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/plan-brainstorm
SKILL.md
Plan Brainstorm
The Planner drives this workflow. Generate creative approaches to a problem using the Diverge → Converge → Synthesize framework. Actively seek external sources (documentation, upstream repos, specifications) to validate and enrich ideas.
Phase 1: Frame the Challenge
Define the problem space before generating ideas.
Steps
- State the problem or design challenge clearly
- Identify constraints that are truly fixed vs. those that can be relaxed
- Define success criteria for the brainstorm output
- Note any "sacred cows" — assumptions worth questioning
Output
Produce a challenge frame covering: challenge statement, fixed constraints, bendable constraints, sacred cows.
Approval Required: No
Phase 2: Diverge
Generate options without judgment.
Steps
- List at least 3 distinct approaches — aim for breadth over depth
- Include at least one unconventional or contrarian option
- Consider "do nothing" as a valid option
- Draw from analogies in other domains, projects, or ecosystems
- For each option, seek external inspiration:
- Use
WebFetchto read how others solved similar problems - Use
mcp__github__search_repositoriesto explore prior art in open source - Cite external sources for ideas that come from outside the codebase
- Use
Rules for Divergence
- No evaluation during this phase — capture everything
- Quantity over quality at this stage
- Each option gets a one-paragraph description with optional external inspiration note
Output
Produce a diverged options list: each option gets a name, one-paragraph description, and optional inspiration source.
Approval Required: No
Phase 3: Converge
Evaluate options against explicit criteria.
Steps
- Define evaluation criteria upfront (cost, complexity, risk, alignment, reversibility)
- Score or rank each option per criterion
- Identify deal-breakers and must-haves
- Note which options can be combined
- Validate key assumptions with external sources where possible
Output
Produce a convergence evaluation covering: criteria table (criterion | weight | description), option scoring table, combinable options, deal-breakers.
Approval Required: No
Phase 4: Synthesize
Produce a clear recommendation.
Steps
- State the recommended approach with clear rationale
- Document rejected alternatives and why
- Identify risks and mitigations
- Define the decision's reversibility
- Note if this warrants an ADR (use
/skill:plan-adrif so)
Output
Produce a synthesis covering: recommended approach with rationale, rejected alternatives table (alternative | reason rejected), risks table (risk | likelihood | mitigation), reversibility assessment, external sources consulted table.
⛔ CHECKPOINT
STOP. Do not proceed until human approves or redirects the recommendation.
Approval Required: Yes
Error Handling
| Error | Recovery |
|---|---|
| Too few options generated | Expand search to adjacent domains, question constraints |
| All options have deal-breakers | Revisit constraints with human, explore hybrid approaches |
| External sources disagree | Document the disagreement, present both perspectives |
| No clear winner | Present top 2 with explicit trade-off, ask human to decide |
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?